Traffic Safety Evaluation At Signalized Intersections Based On Conflict Between Motor Vehicles And Non-motor Vehicles

With the increase in motorised and non-motorised vehicle ownership and the widespread popularity of electric bicycles,the number of motor vehicle and nonmotor vehicle conflicts at signal intersection has increased significantly in recent years.Traffic conflicts not only affect road traffic capacity,but also affects the safety of life and property of drivers and riders.The study of the causes and solutions of traffic conflicts can help to improve traffic safety and reduce the occurrence of traffic accidents.The thesis selects a typical cross signal intersection in Shanghai as the basis for investigation and research.From the survey data,the basic traffic data of the signal intersection is extracted,such as the individual characteristics and behavioural characteristics of both the motorised and non-motorised vehicles,the road environment and traffic flow environment data of the signal intersection.The videos are pre-processed to identify traffic conflicts,determine the basic sample size and record the number of motor vehicle and non-motor vehicle conflicts,and obtain the number of motor vehicle and non-motor vehicle conflicts and conflict-related characteristics data for each intersection.By summarising the quantitative indicators of different types of traffic conflicts,one indicator is selected from each of the three categories of quantitative indicators of traffic conflicts as an indicator for analysing the severity of motor vehicle and nonmotor vehicle conflicts,and the trajectory of the surveyed samples is tracked by means of video detection,and the specific values of the selected three categories of traffic conflict indicators are calculated by applying the traffic conflict indicator calculation model.The conflict severity of the survey sample was classified by Kmeans mean clustering.Twenty-two influencing factors were selected as model variables from four aspects,and a stochastic parametric logit regression model was built to analyse the degree of influence of each influencing factor,A random parametric Logit regression model was established to complete the analysis of the influencing factors for each type of motor vehicle and non-motor vehicle conflicts,to analyse the factors that have a significant influence on the three different severity levels of motor vehicle and nonmotor vehicle conflicts,and to calculate the average marginal benefit of each significant influencing factor to quantitatively investigate the influence of each variable on the severity of the three types of motor vehicle and non-motor vehicle conflicts.Based on the results of the random parameter Logit regression model analysis,12 motor vehicle and non-motor vehicle safety evaluation indicators were selected to construct an ex-ante evaluation model of the motor vehicle and non-motor vehicle conflicts at signal intersections,and the traffic conflicts they cause,to establish a presafety evaluation system for motor vehicles and non-motor vehicles at signal intersections,and to construct a signal intersection safety evaluation system using the entropy-weighted Topsis method The intersections are evaluated using the entropyweighted Topsis method,and the intersections are graded in terms of their safety.Based on the results of the pre-safety evaluation model,the causes of motor vehicle and non-motor vehicle conflicts at each intersection are summarised and targeted safety improvement suggestions are put forward.